The danger of over prescriptive requirements

[ChrisF323949498]

Have you ever been asked: ‘I need a dashboard to show me how many Changes there are in ServiceNow, overtime and I need it to update every 2 minutes, oh and I want the graph to be pink’.

The requirements start to get quite specific… you should pause here.

Every 2 minutes? For Change records? Hmm. A dashboard (not a report?)? Pink?

Prescriptive requirements can be good, for example if we’re baking cookies then knowing how much flour we need is pretty important (and figuring that out by trial and error will give us a real dry mouth, and probably waste a lot of flour)

Generally, however, in the ServiceNow space as a consultant your position will typically be: ‘Given a high level problem, find the best solution’

And that’s where the above goes slightly off-piste.

They aren’t high level, we’ve already starting talk about colours, counts of records, and how often it should update.

So take a pause.

Ask the requester to provide a higher level problem statement.

What is the underlying concern or issue here?

What will this report give them? Why? Why? Why? (Ok not too many whys, that'll probably get annoying)

‘As a Lead, I really need to get visibility into how we’re performing as a Change team because I’m worried we’re over worked recently and falling behind, and I need to decide whether we need to hire more staff or something…help!’

OK, cool, now we get a better feel for the issues here.

It's more a case of their team is swamped and they're ultimately looking for ways to solve what seems to be a capacity problem.

The Dashboard was just another data source to assist them (and it may still be valuable, but there are a whole other list of possible options too now…).

Now we can ask other questions, and there are plenty but some may be…

• Where has this concern come from, why now, why not 2 months ago, what’s happened recently to trigger this?
• Are we as an organisation raising too many changes?
• Can we simplify and automate?
• What’s good performance even look like?
• What metrics can we collate to determine current base line performance?

Capacity issues are input vs output, can we reduce the input? Can we automate the output? Etc.

There’s a lot more to it and I won’t aim to provide an exhaustive list, but the general take away is:

‘If you’re given very prescriptive requirements, pause, ask what the underlying high level problem is’.

Otherwise you may find yourself building a Dashboard, which delivers only 0.01% value vs solving the actual problem.

This is just my 2cents, what’s been your experience?
What helped? What hindered?
 

Cheers all!